Title the COST of MARRIAGES and MARRIAGE-RELATED

Home , Engagement

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Kyoto University Research Information Repository

THE COST OF MARRIAGES AND MARRIAGE-RELATED Title SAVING IN JAPAN

Author(s) Horioka, Charles Yuji

Citation Kyoto University Economic Review (1987), 57(1): 47-58

Issue Date 1987-04

URL http://hdl.handle.net/2433/125573

Right

Type Departmental Bulletin Paper

Textversion publisher

Kyoto University ISSN 0023-6055

THE KYOTO UNIVERSITY ECONOMIC REVIEW MEMOIRS OF THE FACULTY OF ECONOMICS KYOTO UNIVERSITY

VOL. LVII, NO. I APRIL 19 B 7 WHOLE NO. 122

CONTENTS

Ethical Base of J. Habermas' Critical Socialtheory Toshihiko HIRAI I

A Multi-Sectoral Econometric-Model Divided by Social Strata of the Japanese Economy Masano.i NOZAWA Hiroshi OHNISHI 15

The Cost of Marriages and Marriage-Related Saving in Japan Charles Yaji HOltlOKA 47

PUBLISHED BY THE FACULTY OF ECONOMICS, KYOTO UNIVERSITY SAKYO·KU, KYOTO, JAPAN THE COST OF MARRIAGES AND MARRIAGE RELATED SAVING IN JAPAN'

By Charles Yuji HORIOKA*'

I Introduction

It is often said that wedding ceremonies are unusually lavish in Japan and that other marriage~related expenses are also greater than in other countries. It is sometimes further argued that the high level of marriage-related expenses makes it necessary for Japanese households to save in advance and that this is one' factor behind Japan's ex· ceptionally high household saving rate. The foregoing assertions are examined in detail in this paper. The organization of the paper is as follows: In Secti6n'I~, data on the level and composition of marriage-related expenses and on the importan.ce of saving fq,r marriage as a motive for saving are presented and analyzed. In Section III, the aggregate ·amount of marriage-related saving is computed. There is a brief concluding section. To anticipate the main findings of the analysis, marriage~related expenses are considerable in Japan, with the cost per couple ranging from 1.2 to 1.6 times average annual household disposable income. Such a high level of inartiage~related expenses makes it necessary for Japanese households to save in advance, and indeed, saving for one's children's marriage expenses is one of the top five motives for saving in Japan. Somewhat surprisingly, 'however, the aggregate amount of marriage~related saving in Japan is negligible (far less than one percent of household disposable income) because, even though the marriage~related saving of households with pre-marriage children is substantial, it is almost fully offset by the marriage-related dissaving of households in which someone marries during the current year.

n The Cost of Marriages

It is fortunate that at least two surveys of marriage-rel&ted expenses are conducted on a regular basis in Japan, one by the Keio Department Store' and one by the Sanwa 1 Bank. ) Both surveys collect information on all niarriage-related expenses including

01< The author is grateful to participants of seminars at Osaka University's Institute of Social and Econo mic Research and the National Bureau of Economic Research, especially Kazuo Sato, for valuable comments, to Mika Fujii for capable research assistance, and to the Keio Department Store and the Sanwa Bank for generously providing data from their surveys on marriage expenses. Any errors that remain are the sale responsibility of the author, and any views expressed herein are those of the author alone . • * Associate Professor, Faculty of Economics, Kyoto University. 1) Edwards {19B7) analyzes earlier data from the same Sanwa Bank survey as well as data obtained from field research at a commercial wedding facility. 48 c.y. HORIOKA not only the cost of the wedding ceremony/reception itself but also the cost of betrothal gifts (yuinO), engagement gifts, gifts for wedding guests, gifts for the go-between (na kodo), the honeymoon trip, fwnishings and appliances for the new home, and other expenses. Data from both surveys on the average cost of a marriage are shown in Table I, and as this table shows, marriage costs in Japan are high by any standard. The average cost of a marriage is 1.2 to 1.4 times average annual household disposable income according to the Keio Department Store survey and 1.4 to 1.6 times average annual household disposable income according to the Sanwa Bank survey. Of course,

Table 1. Trends over Time in Marriage-related Expenses in Japan

Keio Dept. Store Survey Sanwa Bank Survey

Year Amount per Ratio to Amount per Ratio to couple disposable couple disposable (in yen) income (in yen) income

1972 1,827,000 1.20 na n. 1973 2,485,000 1.37 n. n. 1974 3,011,000 1.34 n' na 1975 3,317,000 1.28 na n. 1976 3,637,000 1.30 4,028,000 1.44 1977 3,971,000 1.29 4,288,000 1.39 1978 4,211,000 1.30 4,518,000 1.39 1979 4,344,000 1.26 5,056,000 1.47 1980 4,647,000 1.27 5,745,000 1.57 1981 5,144,000 1.35 6,124,000 1.61 1982 5,526,000 1.37 6,076,000 1.51 1983 5,773,000 1.40 6,312,000 1.53 1984 5,965,000 1.38 6,778,000 1.57 1985 6,099,000 1.36 7,111,000 1.59 1986 6,236,000 1.37 7,479,000 1.64

Notes : na denotes not available. Both marriage expense surveys collect data on a midyear-ta-midyear basis. The data have been converted to a calendar year ba'!is by averaging the figures for con- secutive years. The Keio Department Store survey is conduc.ted only every other year. Figures for the intervening years were interpolated by assuming a constant rate of increase. Data on the average annual disposable income of salaried worker households were used. Sources: Formarriagt' expense data, Keio Department Store (KeiCi Hyakkaten), Keto Biweekly, Buraidaru Tokusho (Special Bridal Issue), various issues. and Sanwa Bank (Sanwa GinkO), Kyoshiki Zeng(l no Suitobo (Account Book of Expenses at or about the Time of the Wedding Ceremony). Sanwa Ginko -H~mu Konsarutanto Chrisa Repoto (Sanwa Bank Home Consultant Research Report). no. 146. September 1987. p. 3. For data on household disposable income, Management and Coordination Agency, Statistics Bureau (SCimu-cM. TCikei-kyoku). Kakei ChOsa Nenpo (Annual Report on the Family Income and Expenditure Survey), 1986 edition (Tokyo: Zaidan H~jin Nihon TCikei KYCikai [Japan Statistical Association], 1987), p. 40. TIlE COST OF MARRIAGES AND MARRIAGE-RELATED SAVING IN JAPAN 49 this amount is borne jointl-y by the bride's family and the,groom's-family, but even so, each side must put up 0.6 to 0.8 times its annual disposable income.- It would be next to impossible to finance an amount of this magnitude out of current income, and thus it appears virtually certain that Japanese households save in advance for marriage expenses. Table 2 presents data on the breakdown of marriage-related expenses by category, and as the table shows, marriage-related expenses can be grouped into those relating to the engagement, those relating to the wedding, those relating to the honeymoon trip, and those relating to setting up the new household, with the four categories accounting

Table 2. The Composition of Marriage-related Expenses, 1985-86

Category In yen In U.S. Percent dollars of total

Betrothal gifts (yuini5) ¥732,000 $ 4,BBO Engagement gifts 494,000 3,293 6.7% Miscellaneous engagement.related expemes 170,000 1,133 2.3

Wedding ceremony and reception 2,118,000 14,120 2B_B Wedding ceremony 79,000 527 ( U) Food/drinks at reception 1,113,100 7,540 (15.4) Wedding costumes 484,000 3,227 ( 6.6) Other expenses (gifts for wedding guests, photography costs, etc.) 424,000 2,827 ( 5.B) Gift for go-between (nakildo) 120,000 BOO 1.6 Miscellaneous (gratuities, trans- portation costs, etc.) 227,000 1,513 3.1

Honeymoon trip 1,026,000 6,B4{) 14_0

Furniture, electrical appliances, kitchen goods, etc. I,B06,ooO 12,040 24.6 Kimonos, Western-style clothes, 5,653 11.5 accessories, etc. 848,000 Housing (if newly rented) 535,000 3,567 7.3

Total expenses 8,076,000 53,B4{) Betrothal gifts 732,000 4,BBO Total actual expenses 7,344,000 48,960 100.0

Notes: The conversion to U.S. dollars was done using an exchange rate of 150 yen per dollar. The percentage figures take total actual expenses (net of betrothal gifts) as the base. Betrothal gifts are cash gifts from the groom's family to the bride's family and vice versa and hence do not constitute a net expense from the point of view of the two families collectively. Source: Sanwa Bank (Sanwa Ginko), Kyoshiki Zengo no Suitobo (Account Book of Expenses at or about the Time of the Wedding Ceremony), Sanwa GinkG HGmu Konsarutanto Chi5sa Repi5to (Sanwa Bank Home Consultant Research Report), no. 138, September 19B6, p_ 3_ 50 c.y. HORIOKA for 9.0, 33.6, 14.0, and 43.4 percent of the total, respectively. The largest engagement rellilted expense is the engagement gifts (6.7 percent); betrothal gifts are not a. true ex -p~nse because they,are merely transfers of cash fro,m one side to the other. The largest wedding-related expe!>ses are the reception (15.4 percent) and the wedding costumes (6.6 percent), while household formation expenses include furniture, electrical applian ees, kitchen goods, etc., for the new home (24.6 percent), clothes and accessories (11.5 percent), and housing (if newly rented) (7.3 percent). These data provide some insights- into the reasons why marriage-related _expenses are so high in the case of Japan. First, there are a number of expenses that are neces sary in the case of Japan but not necessary in the case of many other countries (for example, multiple weddIng costumes [ir~naoshi], gifts for the wedding ,guests, gifts for for the go-between, etc.). As another example, the cost of setting up a new home is incurred more often in the case of Japan. The reason is that, in Japan, it is common for both the bride and the groom to live with their parents (or in a company dormitory) until marriage and to set up an independent household for the first time at marriage, whereas in many other countries, the bride and/or the groom often live independently of their parents even before marriage, as a resuIt.ofwhich there is no need to set up a new household-at marriage (the 'bride can merely move into the groom's home or vice versa). Moreover, not only is it more often necessary to set up a new h~me in the 'case of Japan but the cost of doing so is,also higher. For example, up to six m,:mths' rent IntiSt be paid in a lump sum when renting an apartment to cover gift money for the 'landlord, the seclll'ity deposit, the realtor's fee, and the first month's rent, and in addition, since furnished' apartments are virtually unavailable in Japan, unlike elsewhere, a full set of furniture, electrical appliances, etc., must generally be proclll"ed. Second, even expenses that are incurred in other countries as well (such as en gagement gifts, the wedding reception, and the honeymoon trip) appear to be much more costly in the case of Japan. Thus, the exceptionally high level of marriage-related expenses in Japan is ap parently due to a number of factors. Turning next to the way in which these expenses are financed, Table 3 presents data on this point, and as the table shows, the most important source of funds is the savings of the bride and of the groom, ~hich account for 47.1 percent of the total. The second most important source of funds is the parents of the bride and groom, with assistance (outright gifts) and loans from the parents accounting for 38.2 and 2.4 petcent of the total, respectively. The parents thus provide 40.6 percent of the total in one form or another, and the bulk of this amount presumably comes from the parents' savings. Hence, the savings of the bride, the groom, and their parents collectively account for a full 87.-7 percent of-the total, with the remaining 12.3 percent being financed by loans from those other than the' parents, wedding gifts, and other sources. This finding, confirms my earlier conjecture that Japanese households must, saVe in 'advance in order to meet their marriage-related expenses. Further evidence on the importance of saving for marriage-reJ::lted expenses is provided in Table 4, which presents data collected by the Central Council for Savings THE COST OF MARRIAGES AND MARRIAGE-RELATED SAVING IN JAPAN 51

Table 3. The Fi,nancing of Marriage-related Expenses, 1985-86

Source of funds Amount Amount in Percent of in yen U.S. dollars total

The bride and groom's own savings ¥3,802,000 $ 25,347 47.1% Assistance from parents 3,085,000 20,567 38.2 Loans from parents 195,000 1,300 2.4 Loans from those other than parents 414,000 2,760 5.1 Wedding gifts, etc. 580,000 3,867 7.2

Total expenses 8,076,000 53,840 100.0 Betrothal gifts received. -732,000 -4,880

Total actual expenses 7,344,000 48,960

Notes: The conversion to U.S. dollars was done using an exchange rate of 150 yen per dollar. The percentage figures take total expenses (gross of betrothal gifts) as the base. Source: Sanwa Bank (Sanwa Ginkci), Kyoshiki. Zengo no Suitobo (Account Book of Expenses at-or about the Time of the Wedding Ceremony), Sanwa Ginkl5 Hi5rnu Konsarutanto Ch1lsa Repato (Sanwa Bank Home Consultant Research Report). no. 138, Septem~ ber 1986, p. 3.

Table 4. Saving Motives

Year Motive 1984 1985 1986

For illness and other unforeseen disasters 75.0 77.2 75.0 34.4 31.4 31.6 For one's children's educational expenses 41.6 43.0 44.7 13.8 14.6 14.3 For one's children's marriage expenses 17.6 17.1 15.3 4.5 3.5 3.3 For the purchase, expansion, or renovation of land/housing 26.3 19.8 20.5 12.3 9.0 9.9 For living expenses during old age 42.1 42.5 42.5 15.5 16.6 15.9 For the purchase of consumer durables 7.5 10.5 10.8 1.1 0.9 0.8 For leisure expenses 9.7 4.8 5.2 1.0 0.5 0.3 For the payment of taxes 5.2 5.4 5.5 0.9 0.7 0.5 No specific motive but for peace of mind 25.7 26.4 25.3 6.5 6.6 7.1 Other 1.2 1.6 1.4 0.7 1.0 0.9

Notes The top figure shows the percent of respondents for whom each motive is one of their top three motives for saving, while the bottom figure shows the percent of respon~ dents for whom each motive is their single most important motive for saving. Source: Central Council for Savings Promotion (Chochiku ZCikyCi Chnr>" Iinkai), Chochiku ni kansuTu SeTon Chosa (Public Opinion Survey on Saving), 1986 edition, Tables 8-9, p. 112. 52 c.y. HORIOKA

Promotion on the percent of households saving for each motive. The motives "for illness and other unforeseen disasters" and "'no specific motive but for peace of mind" are both essentially precautionary motives, and jf the two are considered collectively, they constitute by far the most important motive for saving. "For one's children's educational' expenses" and "for living expenses during old age" are roughly tied for second place and are followed by "for the purchase, expansion, or renovation of landJ housing" in fourth place and "for one's children's marriage expenses" in fifth place. Thus, saving for marriage·related expenses is one of the top five motives for saving in Japan with one in six households counting this motive as one of their top three motives and one in twenty to thirty households counting it as their single most important motive. Note, moreover, that these figures may be underestimates because the marriage·related category in the questionnaire includes only saving for one's children's marriage expenses

Table 5. Saving Targets by Motive

The av~rage target The rcent of The aggregat~ target permobveforhouse- pe for each motlve as a Motive holds saving for households saving percent of the aggre each motive (thou- for each motive gate target for all sands of yen) motives Old age 3389(1) 42.5(3) 53.5(1) Future business capital 3089(2) na na Purchase of land/housing 2550(3) 20.5(4) 19.4(2) Children's education 726(4) 44.7(2) 12.1(3) Children's marriage 614(6) 15.3(5) 3.5(5) Unexpected expenditures 363(7) 75.0(1) 10.1(4) Purchase of durables 192(8) 10.8(6) 0.8(6) Recreation 136(9) 5.2(8) 0.3(8) Payment of taxes na 5.5(7) na Other 662(5) 1.4(9) 0.3(7)

Notes : na denotes not available. The figures in parentheses denote the rank of each motive. Sources: The target data in the first column are from the "Shi5hisha no Seikatsu Ishiki to Shr;hi Chochiku K(kii5 ni kansuru Jittai Chi5sa" (Survey concerning the Attitudes and Consumption/Saving Behavior of Consumers), which was conducted by the Research Institute on National Life (Kokumin Seikatsu Kenkyiisho) in November 1966. 551 regular households living in the 23-ward area of Tokyo, Musashino City, and Mitaka City were surveyed. The data are presented in Research In stitute on National Life (Kokumin Seikatsu Kenkyilsho), ed., Kokumin Seikatsu Tokei Nenpo (Statistical Yearbook on National Life), 1969 edition (Tokyo: Shiseid15, 1969), Table 160, p. 164. The data on the percent of households saving for each motive (second column) are reproduced from Table 4. The figures are for 1986 and denote the percent of households for which each motive is one of the three most important motives for saving. The figures in the third column were computed from the figures in the first two colunms as follows: the product of the fus~ two cohunns was calculated and scaled so that the sum of the products would equal 100. '!HE COST OF MARRIAGES AND MARRIAGE-RELATED SAVING IN JAPAN 53 and excludes saving for one's own marriage expenses, even though we found the latter to be the most important source of funds for marriage-related expenses. Still further evidence on the importance of saving for marriage-related expenses is given in Table 5. The first column of this table presents data on saving targets by motive from a 1966 survey conducted by the Research Institute on National Life, and as the figures show, the saving target for one's children's marriage expenses is the sixth largest (behind old age, future business capital, purchase of land/housing, children's education, and other, in that order). Finally, the third column of the same table presents data on the aggregate target for each motive as a percent of the aggregate target for all motives and shows that the aggregate target for marriage is the fifth largest, accounting for 3.5 percent of the ag gregate target for all motives. In sum, the foregoing evidence has shown that marriage-related expenses are exceptionally high in Japan, with the cost per couple ranging from 1.2 to 1.6 times average annual household disposable income. Such a high level of marriage-related expenses makes it necessary for Japanese households to save in advance, and indeed, saving for one's children's marriage expenses is one of the top five or six motives for saving in Japan. The problem with the analysis thus far, however, is that it has neglected to take account of the dissaving for each motive. In the case of all motives other than the precautionary motive, households save with the intention of eventually drawing down their savings in order to realize the motive in question. For example, households with pre-marriage children save for their children's marriage expenses until the date of the marriage, at which time they draw down those savings in order to meet their children's marriage expenses. Furthermore, at any given time, there will be those saving for a particular motive as well as those dissaving in order to realize that same motive, and the saving of the former will be partly or fully offset by the dissaving of the latter, with the offset being complete in the case of a stationary economy. Thus, no matter how large is the amount of saving for a particular motive, its contribution to the aggregate household saving rate will not necessarily be large and may even be zero or negative, depending on the relative amounts of saving and dissaving for that motive. In the following section, I will calculate the aggregate amount of marriage-related saving taking full account of these considerations.

m The Aggregate Amount of Marriage-related Saving

As argued above, the aggregate amount of (net) marriage-related saving can be calculated as the (gross) marriage-related saving of households with pre-marriage children minus the marriage-related dissaving of households in which someone- marries during the current year. The amount of marriage-related dissaving can easily be calcu lated since detailed data on actual marriage-related expenses are available, but un fortunately, direct data on marriage-related saving are not available. I have therefore 54 c.Y. HORIOKA had to resort to computing the amount of marriage-related saving indirectly. I have done so by assuming that households begin saving for a given child's marriage expenses a. certain period of time before the date of the marriage and that they save at the rate needed to accumulate the necessary amount by the date of the marriage. In particular, I have assumed that the household begins saving for a given child's marriage four years before the year in which the marriage takes place and that it saves two-ninths of the present ·value of its share of the total cost of the marriage in each year, financing the remaining one-ninth out of current income. The assumption about financing one ninth out of current income is reasonable because marriages take place in June, on average, meaning that the household has only six months in which to save during the year in which the marriage takes place and can therefore be expected to save only one half of the amount saved in previous years. The only other assumption requiring justification is the assumption that households begin saving four years before the marriage. There are at least four justifications that can be given for this assumption, to wit: (I) According to Ministry of Health and Welfare (1987, p. 67, p. 385), there were 735,850 marriages in 1985 and a population of 120,265,700. Thus, there were 0.61 marriages per 100 people, meaning that 1.22 percent of the population was married in that year. Moreover, according to Management and Coordination Agency (1986a, p. 38), the average household size in 1985 was 3.71 persons, meaning that a marriage took place in 4.54 percent of households in 1985. Next, recall from Table 4 that 17.1 percent of households were saving for their children's marriage expenses in 1985. Since the percent of households saving for marriage is 17.1/4.54=3.77 times greater than the percent of households in which a marriage took place, we can deduce that Japanese households begin saving for their children's marriage expenses 3.77 years before the marriage. (2) It can be calculated from the data in Ministry of Health and Welfare (1987, pp. 98, 358--359) that the father is on average about 55 to 59 years old when his children marry. Moreover, according to Central Council for Savings Promotion (1985, pp. 40--41), the percent of households saving for their children's marriage expenses is by far the highest for households in which the head is in his fifties. Thus, Japanese households apparently do not begin saving for their children's marriage expenses until just before the marriage. For example, if we assume that households begin saving for marriage when the head is 55, that is only zero to four years before the marriage. (3) According to Ministry of Health and Welfare (1987, p. 359), the mean age at first marriage in 1985 was 28.2 years for males and 25.5 years for females. If we bear in mind that the average age of graduation from college is 22 to 23 in Japan and that Japanese parents are unlikely to be able to begin saving for their children's marriage expenses before their children graduate from coJlege because of the burden of college related expenses, we can surmise that Japanese parents (at least the parents of college graduates) begin saving for their children's marriage expenses at most five to six years beforehand in the case of sons and 2.5 to 3.5 years beforehand in the case of daughters, THE COST OF MARRIAGES AND MARRIAGE-RELATED SAVING IN JAPAN 55

even if they have only one child, and even later if they have more than one child. Moreover, since parents are unlikely to be able to begin saving for the marriage expenses of younger children (or those marrying later) before their older children (Of_ those marrying earlier) marry, this will further delay the start of marriage-related saving, at least for the younger children (or those marrying later). (4) We saw earlier that each side must contribute about 0.6 to 0.8 times 'average annual household disposable income toward the cost of each child's marriage. Assum ing that the most a household can comfortably afford to save for a given motive is fifteen percent of its disposable income, both sides would have to begin saving about four years before the marriage. The above considerations suggest that our best estimate o(when households begin saving for marriage expenses is four (or less) years beforehand. Thus, the assumption I have made appears to be a reasonable one. I am now ready to discuss the specific calculation method used. Gross saving, dissaving, and net saving for marriage were calculated as follows:

. 2 ' [N .w .] 2 4 [ IT, (\ +g,+j)] (I) Gross savmg GSt =- ~ j 1+1 t+1 =- ~ N t +iWt"'7r'----- 9 i.' II (I +r/+j) 9 i_' II (I +r,+j) j_l i-1 (2) Dissaving DS, ~ : N,W,

(3) Net saving NS, ~GS,-DS" where the variables were defined as follows: N =the number of marriages W =marriage~related expenses per couple g ~the growth rate of W r=the interest rate. These formulas reflect the assumptions discussed earlier and therefore do not require further explanation except on one point: Dissaving is calculated as eight~ninths of marriage expenses incurred during the current year because -we assumed that eight~ ninths of the total amount is financed out of past savings and the remaining one~ninth out of current income. Note that the formulas confirm that the amount of gross marriage-related saving will equal the amount of marriage-related dissaving, meaning that net marriage~related saving will be zero, in a stationary economy in which the number of marriages is cons~ tant and the growth rate of maITiage~related expenses equals the interest rate. If, however, the number of marriages is increasing (either because the pop-y.lation is in~ creasing or because the marriage rate is increasing, due for example to an increase in the proportion of the population that is of marriageable age) or the growth rate of marriage-related expenses exceeds the interest rate, the amount of gross marriage~ related saving will exceed the amount of marriage-related dissaving, leading to a positive amount of net marriage-related saving. ' Turning next to a discussion of the data used'in the calculatiohs,'-refer to'the notes 56 C.Y. HORIOKA

2 to Table 1 for the source of the data on marriage.related expenses. ) Data on the interest rate were taken from Management and Coordination Agency (1986b, Table 12-25, pp. 418-419) and the corresponding table of earlier editions of the same; the daily average of the interest rate on one-year term deposits at banks from July I of the previous year to June 30 of the current year was used. Finally, data on the number of marriages were taken from Ministry of Health and Welfare (1987, Table 3-2, pp. 66-67).

Table 6. Trends over Time in the Aggregate Amount of Marriage-related Saving

Keio Dept. Store Survey Sanwa Bank Survey Year Gross Net Gross Net Dissaving Dissaving Saving Saving Saving Saving

1972 3.82 3.00 0.82 22.23 17.45 4.78 1973 3.22 3.23 -0.01 17.28 17.35 -0.07 1974 2.66 2.94 -0.28 12.75 14.11 -1.36 1975 2.37 2.61 -0.24 11.26 12.43 -1.16 1976 2.18 2.34 -0.15 2.48 2.59 -0.11 10.23 10.96 -0.73 11.63 12.13 -0.50 1977 2.11 2.19 -0.08 2.46 2.37 0.09 10.61 11.03 -0.42 12.36 11.91 0.45 1978 2.05 2.06 -0.01 2.40 2.21 0.19 10.81 10.88 -0.07 12.67 11.68 0.99 1979 2.02 1.99 0.03 2.33 2.32 om 12.78 12.59 0.19 14.75 14.65 0.09 1980 1.98 1.94 0.04 2.24 2.40 -0.17 13.04 12.78 0.26 14.71 15.80 -1.09 .

Notes: The upper figure shows marriage-related saving as a percent of household disposable income, while the lower figure shows marriage-related saving as a percent of household saving. Refer to the text for a description of the calculation method and data source for marriage-related saving. Data on household disposable income and household saving were taken from Econo mic Planning Agency (Keizai Kikaku-cho), ed., ShOwa 55-nen Kijun KaUri Kokumin Keizai Keisan Hoko/cu (Report on Revised National Accounts on the basis of 1980), volume 1 (Tokyo: Okura-shi:S Insatsu-kyoku [Ministry of Finance Printing Bureau], 1986), Account 5, pp. 108--111, and Economic Planning Agency (Keizai Kikaku-cho), ed., Kokumin Keizai Keisan Nenpo (Annual Report on National Accounts), 1987 edition (Tokyo: Okura-shi:S Insatsu-kyoku [Ministry of Finance Printing Bureau], 1987), Account 5, pp. 102-103. However, following Hayashi (1986), the consumption of fixed capital, which is valued at historical cost in the above ~soun,es, was re-valued at replacement cost.

2) Note that the data on marriage-related expenses represent total expenditures thereon and cannot be broken down into price and quantity components. Thus, increases therein reflect both in creases in price and increases in quantity. THE COST OF MARRIAGES AND MARRIAGE-RELATED SAVING IN JAPAN 57

Data on household expectations concerning the date of marriage, the growth fate of marriage expenses, and the interest rate were needed to compute the marriage-related saving of households with pre-marriage children. I have adopted the assumption that households have perfect foresight about all of these variables. I also tried other expectations mechanisms but found that the results were not significantly affected. Table 6 presents the results of the calculations, and as the table shows, the amount of gross marriage-related saving is substantial, ranging from two to four percent of household disposable income and ten to twenty-two percent of household saving, but is offset by almost as large an amount of marriage-related dissaving (two to three percent of household disposable income and ten to seventeen percent of household saving). Consequently, the amount of net marriage-related saving is negligible and often negative, ranging from -0.3 to 0.04 percent of household disposable income and -1.4 to 1.0 percent of household saving if the year 1972 is excluded. The amount of net marriage related saving was unusually large in 1972 (0.8 percent of household disposable income and 4.8 percent of household saving), but this was because the growth rate of marriage expenses was exceptionally high in the subsequent two years (36 percent in 1972-73 and 21 percent in 1973-74). As for why the amount of net marriage-related saving was negligible and often negative after 1972, there are two reasons. First, despite the continued growth of the population, the number of marriages has beer;t declining steadily since 1972, when the number of marriages of the postwar baby boom generation (born in 1947-49) reached its peak. Second, the growth rate of marriage expenses has stabilized at roughly the same level as the interest rate and was lower than the interest rate in certain years. The amount of net marriage-related saving must have been greater before 1972 because the number of marriages increased steadily from 1951 until 1972 and because the growth rate of marriage expenses probably exceeded the interest rate,3) but an exact calculation -could not be done due to the unavailability of data on marriage expenses prior to 1972. I was able to perform a rough calculation, however, which showed that the amount of net marriage-related saving was far less than one percent of household disposable income and five percent of household saving in every case.

IV Conclusion

The analysis in the previous section showed that, despite the importance of mar riage expenses and marriage-related saving from the point of view of individual house holds, the aggregate amount of net marriage-related saving has been negligible, ex plaining far less than one percentage point of Japan's household saving rate, which has recently averaged fifteen to twenty percent.

3) Edwards (1987) describes the rapid rise of the commercialized wedding industry and the attendant increase in wedding costs during the 1960s and early 1970s. He suggests that improved living standards are one factor behind this trend. a reasonable assertion ,since wedding services are presumably a luxury good with an income-elastic demand. 58 C.Y. HORIOKA

Moreover, I have conducted similar analyses of education-related saving and housing-related saving (Horioka (1985) and Horioka (forthcoming)) and have found that they, too, are important from the point of view of individual households but un important at the aggregate level, explaining less than one and les~ than two to three percentage points, respectively, of Japan's household saving rate. Thus, saving for these-three motives collectively accounts for less than five percentage points of Japan's household saving rate. My conjecture is that the remaining ten to fifteen percentage points is primarily saving for old age, bequests, and precautionary purposes, and I therefore hope to analyze saving for these motives as part of my future research agenda.

Bibliography 1. Central Council for Savings Promotion (Chochiku ZlJkylJ Choe linkai) (1985), Chochiku ni kansuru Seron ChOsa (Public Opinion Survey on Saving), 1985 edition. Tokyo: Chochiku ZlJkylJ Chni5 Iin kai. 2. Central Council for Savings Promotion (Chochiku Zi5ky5 Choo linkai) (1986), Chochiku ni kansuru Seron ChOsa (Public Opinion Survey on Saving), 1986 edition. Tokyo: Chochiku ZokylJ Choo lin kai. 3. Economic Planning Agency (Keizai Kikaku-chCi) (1986), ed., ShOwa 55-nen Kijun Kaitei Kokumin Kei zai Keisan Hokoku (Report on Revised National Accounts on the Basis of 1980), volume 1. Tokyo: Okura-shlJ Insatsu-kyoku (Ministry of Finance Printing Bureau). 4. Economic Planning Agency (Keizai Kikaku-cho) (1987), ed., Komin Keizai Keisan Nenpa (Annual Report on National Accounts), 1987 edition. Tokyo: Okura-she; Insatsu-kyoku (Ministry of Finance Printing Bureau). 5. Edwards, Walter (1987). "The Commercialized Wedding as Ritual: A Window on Social Values," Journal of JapaTUIseStudies, volume 13, number 1 (Winter), pp. 51-78. 6. Hayashi, Fumio (1986), "Why Is Japan's Saving Rate So Apparently High?" in Stanley Fischer, ed., NBER Macroeconomics Annual 1986 (Cambridge, Massachusetts: The MIT Press), pp. 147-210. 7. Horioka, Charles Yuji (1985), "The Importance of Saving for Education in Japan," The Kyoto University Economic Review, volume 55, number 1 (April), pp. 41-78. 8. Horioka, Charles Yuji (forthcoming), "The Importance of Saving for Housing Purchase in Japan." Osaka University, Institute of Sodal and Economic Research, Ibaraki, Osaka, Japan. 9. Keio Department Store (KeiCi Hyakkaten) (1987), Keio Biweekly, Buraidaru Tokushn (Special Bridal Issue). Also earlier editions of the same. 10. Management and Coonlination Agency, Statistics Bureau (S!")mu-ch!"), Tl:Skei-kyoku) (1986a), Kakei Chasa Nenpo (Annual Report on the Family Income and Expenditure Survey), 1985 edition. Tokyo: Zaidan Hi5jin Nihon T!")kei KyCikai (Japan Statistical Association). 11. Management and Coordination Agenc'y, Statistics Bureau (Sl:Smu-ch!"), T5kei-kyoku) (1986b), Nihon Takei Nenkan (Japan Statistical Yearbook), 1986 edition (36th edition). Tokyo: Zaidan H!")jin Nihon T!")kei KylJkai (Japan Statistical Association) and Mainichi Shinbunsha (The Mai nichi Newspapers). Also earlier editions of the same. 12. Ministry of Health and Welfare, Minister's Secretariat, Statistics and Information Department (Kosei-shi5, Daijin-kanb!,,), TCikei j!")h!")-bu) (1987), Jinko Dotai Takei (Vital Statistics, Japan), 1985 edition, volume 1. Tokyo: Zaidan Hl:Sjin Kl:Ssei TlJkei Ky!")kai (Welfare Statistics Association). 13. Research Institute on National Life (Kokumin Seikatsu Kenkynsho) (1969), ed., Kokumin Seikatsu Takei Nenpo (Statistical Yearbook on National Life), 1969 edition. Tokyo: Shiseid5. 14. Sanwa Gink!") (Sanwa Bank) (1986), Kyoshiki Zengo no SuitObo (Account Book of Expenses at or about the Time of the Wedding Ceremony), Sanwa Gink!") Hi5mu Konsarutanto Chasa Repato (Sanwa Bank Home Consultant Research Report), no. 138 (September).